Neural Network Model Selection for Dynamic Strain Measurement using Frequency-Domain Parameters of Fiber Optic Sensor

The present study is aimed to develop a smart strain prediction model using fibre optic sensors and neural network. Frequency domain optical parameters are obtained experimentally using a cantilever beam structure, under dynamic loading conditions. Four variations with external damage are used to study strain variations on healthy, single damage and multiple damage beam structures. The strain values were correlated to the set of phase difference, change in real part and amplitude by using feed-forward back propagation neural network approach. The strain values using optical parameters were verified with analytical strain measurement. As the signals always get affected due to the presence of noise, this drawback is eliminated using a well trained neural network model. The neural network model provides a better and advanced methodology for strain prediction compared to the conventional analytical solution.